属性相关选择和AdaBoost算法在入侵检测中的应用

入侵事件的识别是入侵检测系统的关键，入侵事件的识别是一个网络数据的分类问题。通过基于相关的属性选择算法，选择出相关度高的属性子集，去除冗余度高的属性，在选择的属性子集上，使用AdaBoost算法对网络数据分类，识别入侵事件。实验结果表明，在选用的实验数据上，基于相关的属性选择算法和AdaBoost算法结合使用，提高了分类正确率和入侵事件的检出率，降低了入侵事件的误报率。

Correlation-base feature selector and AdaBoost applied in intrusion detection

Intrusion event recognition is the key to intrusion detection systems, and it also is a networkdata classification problems. For high recognition rate,throughthe selection algorithm based on relevantattributes, this paper selects a subset of the attributes with low redundant, identification intrusion byevent AdaBoost algorithm. The experiments show that the correlation-based attribute selection algorithmand AdaBoost algorithm improve the classification accuracy and intrusion detection rate of events,reducing the false alarm rate of intrusion events.